Assessment of PD-1 expression in human resting and activated natural killer cells and murine tumor models.

Abstract

36 Background: Blockade of the PD-1/PD-L1/2 axis has revolutionized cancer therapy. Although reinvigorated PD-1+ T cells are the main effectors in the response to checkpoint blockade, the contribution of Natural Killer (NK) cells to PD-1/PD-L1 inhibition is under debate. While PD-1 has been identified on NK cells, this appears to be restricted to small populations under limited conditions. We sought to evaluate the extent of PD-1 expression in mouse and human resting and activated NK cells. Methods: Human NK cells were isolated from healthy donor PBMCs and cancer patients. Ex vivo activation and proliferation techniques included recombinant human cytokine and feeder line co-culture. Murine NK cells were isolated from splenocytes, and PBMCs from wild type and immunodeficient mice. We assessed NK cell surface markers and intracellular cytokine by flow cytometry, and gene expression by quantitative RT-PCR. Results: Over 21-days of ex vivo expansion, expression of PD-1 or PD-L1 on human NK cells was < 1% at all time points, while TIGIT+ expression increased to > 85%. Conversely, ConA stimulation of T cells increased PD-1 expression with no change in TIGIT expression. QRT-PCR demonstrated absent PD-1 expression in purified NK cells compared to a 5-fold increase in PD-1 gene expression in ConA stimulated PBMCs. PD-1/PD-L1 was also < 1% in the NK92 cell line and < 2.5% in peripheral CD56+CD3- NK cells from patients with soft tissue sarcoma (STS). NK cells from digested freshly resected STS show variable PD-1 ( < 10%) and minimal PD-L1 ( < 1%) expression with a small, but measurable population of intra-tumoral NK cells (1% of immune cells). In vivo mouse studies showed < 5% PD-1+ NK cells in spleen and tumor of CT26 tumor-bearing mice, while PD-L1+ NK cells increased in frequency from spleen (5-35%) to tumor (40-95%) in both wild type BALB/C and SCID mice. Conclusions: In contrast to prior studies, we did not observe a substantial PD-1+ population on human or murine NK cells after multiple activation strategies compared to T cells. Contrary to its application in T cells, our data suggest that PD-1 is not a useful marker for NK cell exhaustion/dysfunction. PD-L1 on NK cells may represent an important link between NK and T cell immunotherapy.